A memory-distributed quasi-Newton solver for nonlinear programming problems with a small number of general constraints

Open Webpage

Cosmin G. Petra

Published: 01 Jan 2019, Last Modified: 11 May 2024J. Parallel Distributed Comput. 2019EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•A data parallel quasi-Newton method for optimization is proposed.•Linear algebra of the method is parallelized for distributed memory computing.•Time and space complexity and theoretical parallel speedup are discussed.•Good parallel speedup on up to 4608<math><mn is="true">4608</mn></math> cores is shown for topology optimization.